Saltzman B. (editor) Anomalous Atmospheric Flows and Blocking
Подождите немного. Документ загружается.
44
RANDALL
M.
DOLE
up but continues to intensify
as
it approaches the key region. Similarly, the
intensification of the main center over the subtropical North Pacific appears
related in part to a disturbance originating in midlatitudes. Concurrent with
these developments, a larger scale pattern of height
rises
over the midlatitude
North Pacific, with falls
to
the south,
is
again evident. Maps following day
0
(not shown) indicate that subsequent developments are qualitatively similar
to those displayed in the low-pass analyses.
As
for the filtered analyses, the unfiltered PAC negative anomaly evolu-
tion (Fig.
5)
in many respects resembles the corresponding positive se-
U
FIG.
5.
Composite time evolution
of
un6ltered
500-mbar
height anomalies (in meters)
ob-
tained
&om
the
same
PAC
negative anomaly
cases
as
in
Fig.
2,
for
days (a)
-
3,
(b)
-
2,
(c)
-
1,
and (d)
0.
Cofiesponding
confidence levels
are
shown
in
a'-d'.
*,p-os
‘OIJ
SP
S3ITVPIONV IN31SISXI3d
JO
S313A3
%I7
quence. The negative center to the east of Japan on day
-
3
in the negative
analyses, however, is considerably more intense than its positive counter-
part. This negative center intensifies and moves eastward to the key region by
day
0;
the second negative anomaly center over the East China Sea on day
-
2
follows
a
similar course, eventually merging
with
the main center after
day
0
(not shown). The difference and corresponding t-test results (Fig.
6)
confirm our expectations based on the separate positive and negative analy-
ses:
immediately in advance of development, the major additional differ-
ences between positive and negative cases not identified in the low-pass
46
RANDALL
M.
DOLE
filtered analyses are associated with the propagating, intensifying midlati-
tude disturbances initially near Japan.
4.3.
Relationship
to
the Zonal
Flow
in the PaciJic Jet Region
The variations in the zonal flow over eastern Asia and the western Pacific
in advance of the developments over the central Pacific show an interesting
FIG.
6.
Composite time evolution
of
unfiltered
500-mbar
height anomaly
differences
(in
meters)
between
the positive
and
negative
anomaly
cases
for
days
(a)
-
3,
(b)
-
2,
(c)-
1,
and
(d)
0.
Corresponding confidence
levels
for
the differences
between
means
are
shown
in
a’-d’.
LIFE
CYCLES
OF
PERSISTENT
ANOMALIES
47
FIG.
6c-d’.
relationship to the climatological mean jet structure, with a tendency for the
largest meridional gradients in the height anomalies to occur over the mean
jet axis. This suggests that the evolving flow anomalies are associated with
major changes in the jet intensity and structure over this region. This rela-
tionship can
be
seen more clearly in Fig.
7,
which displays the time evolution
of the geostrophic wind anomalies for the
PAC
negative cases evaluated
along the wintertime climatological mean jet axis from central Asia to the
eastern North Pacific. We see that well prior to development there is a
modest intensification of the jet well upstream of the key region, with a
weaker than normal jet to the east near the jet exit region. Beginning at about
48
RANDALL
M.
DOLE
ALONG
JET
AXIS
(rn/sec)
-5
L
20-
15
0
0
'5
1
I
I
I
I I
-.
8O"E
100'E
120.E
140-E
160.E
180
170'W
150"W
130"W
30"
30.N
3O9N
35"N 35"N
35'N40°N 40"N 45"N
POSITION
FIG.
7.
Time evolution of the composite geostrophic wind anomalies (in
m
sec-I)
evaluated
along the climatological mean 500-mbarjet axis from eastern Asian to western
North
America
for the
period
from
5
days
before
to
5
days
following onset ofthe Pacific negative anomaly
cases.
day
-3,
the zonal wind anomalies in the western Pacific strengthen and
begin to propagate eastward, reflecting an intensification and eastward shift
of
the jet maxima. Following development, the major flow anomalies re-
main nearly quasi-stationary over the Pacific, reflecting the intensified and
eastward-elongated jet structure. Parallel analyses on the positive cases (not
shown) display a weakening
of
the jet upstream immediately prior to devel-
opment, with pronounced weakening of the jet in the jet exit region over the
central North Pacific during development. Analyses conducted on the At-
lantic cases (not shown) display similar variations in the jet intensity and
structure, with a tendency for a weakening of the upstream jet over eastern
North American and the western Atlantic prior to the development of the
positive cases (which are associated with blocking over the eastern Atlantic)
and an intensification of the jet prior to the negative cases.
4.4.
Representation
of
Development in Zonal Fourier Harmonics
Investigators have suggested that blocking over the oceans is often charac-
terized by the amplification of certain zonal wavenumbers
of
the 5OO-mbar
LIFE
CYCLES
OF
PERSISTENT ANOMALIES
49
a,,,,,,,,,,,,,,,,,
I
'''1
111
I1
1
...
*',,,11...
,I....
-0
-6
-5
-4
-2
0
2
4
6
B
10
12
I*
16
!a
20
I
FIG.
8.
Composite vector time series of zonal Fourier components
k
=
0-6
of the low-pass
anomalies at latitude
45"N
for the
PAC
(a) positive and (b) negative anomaly
cases.
The
amplitude of
a
component is proportional to the length of the vector. The phase is given by the
angle measured
in
the
clockwise direction
between
the positive vertical
axis,
through the origin
of the vector, and the vector itself(positive angles eastward); thus clockwise (counterclockwise)
rotation
with
increasing time indicates eastward
(westward)
propagation. The reference longi-
tude is chosen
as
170"W,
so
that the vertical component of a vector is proportional to the
contribution by that wavenumber
to
the observed anomaly at that longitude and time.
50
RANDALL
M.
DOLE
height fields (Austin,
1980;
Colucci
et
al.,
1981).
We have examined this
possibility for our cases by constructing composite vector time series ofzonal
Fourier components,
k
=
0
-
6,
of the anomalies at the corresponding “key”
latitudes
(Fig.
8).
Consistent with the time mean composites presented in
Dole
(
1985),
we see that important contributions to the observed anomalies
at the key point are provided by a relatively broad band of components,
primarily wavenumbers
k
=
0-4.
Thus, these analyses do not suggest a
strong dominance by a particular wavenumber in the development of the
anomaly patterns.
These analyses also provide no clear indication of a systematic precursor
in the wavenumber domain: prior to onset, the amplitudes are small, typi-
cally under
10
m, and the vectors tend to have a haphazard orientation with
respect to the vertical axis, indicating alack of wave coherence locally. Before
the major amplification, most components show some rotation, indicating
propagation, with a tendency for the lowest wavenumbers
(k
=
1
-2)
to be
retrogressive and the higher wavenumbers
(k
>
5)
to
be
progressive. Dur-
ing the period of maximum amplification, however, little propagation is
evident for wavenumbers
k
<
5,
consistent with an interpretation of this
stage
of
the development in terms of waves having phase speeds near zero.
5.
VERTICAL
AND
THERMAL
EVOLUTIONS
DURING
DEVELOPMENT
A
further indication of the character of the developments is obtained by
examining how the vertical structures evolve in time. For this purpose,
parallel development composites for the
lOOO-, 700-,
500-,
300-,
and
100-
mbar height anomalies were prepared from unfiltered data following the
procedures described previously. Data at all levels were available only for the
11
winters from
1965/1966
through
1975/1976,
so
that the composites in
this section were formed from all of the previous cases [listed in Dole
(
1986)]
that fell within this period. For brevity, only the
PAC
evolutions are dis-
cussed in detail.
Figure
9
presents longitude-pressure cross-sections at
45
ON
and
20”
N
of
the unfiltered
PAC
positive composite anomalies at
1
-day
intervals from day
-
3
to day
0.
We see that, consistent with the previous 500-mbaranalyses, the
rapid development of the main center appears to be partly associated with an
amplifying, eastward-propagating midlatitude disturbance. This feature has
pronounced westward tilts with height during this period, suggesting that a
substantial baroclinic contribution is involved in its amplification. There are
two maxima in the vertical structure, with peaks at
1000
mbar and near
300
mbar.
A
similar structure is observed in numerical studies of nonlinear
effects on growing baroclinic waves of long synoptic scales (Gall,
1976;
LIFE CYCLES
OF
PERSISTENT ANOMALIES
51
P
D
P
100
'
L-hd
'45
1'
230
300
1000
KloO
I20
E
140
160
180
160
140
W
120
120
E
I40
160
b'
180
160
yo
w
eo
b
P
C
0
P
P
120
E
uo
160
160
no
w
120
d'
FIG.
9. Composite longitude-pressure
(p,
millibars)
cross
sections
of
unfiltered anomalies
(in meters)
for
the
PAC positive anomaly
cases
at
45"N
for
days (a)
-
3,
(b)
-
2,
(c)
-
1,
and
(d)
0.
Corresponding
cross
sections
constructed instead at
20"N
are
presented in a'-&.
52
RANDALL
M.
DOLE
Simmons and Hoskins,
1978).
Such a structure, however, may also
be
partly
an artifact of data coverage because, over this region, data are more plentiful
at these levels. Nevertheless, comparison of the relative positions
of
the
centers at the data-rich levels suggests that there are pronounced westward
tilts with height throughout the troposphere, and that, at an early stage in the
development, the maximum anomalies are realized in the upper tropo-
sphere.
The corresponding development at
20"
N indicates that, in parallel with its
midlatitude positive counterpart, the main negative center progresses east-
ward across the Pacific through day
0.
Subtropical negative anomalies are
initially confined primarily to the upper troposphere. Following
day
0
(not
shown), however, 1000-mb heights continue to
fall
over the central Pacific,
leading to the establishment of
a
cold-core negative center with little or no
evidence of tilts throughout the troposphere.
Figure
10
shows simiIar analyses
for
the PAC negative
cases.
In many
respects, the vertical evolution parallels that
of
the positive cases. The main
center propagates eastward and intensifies through the period. The asso-
ciated trough
axis
initially tilts strongly westward with height, but becomes
nearly vertical by
day
0.
Double maxima are also evident in the vertical
structure early in the evolution, giving way to a single major center in the
upper troposphere
by
day
0.
The subtropical patterns are initially rather
illdefined, but a positive center is evident over the subtropical mid-Pacific
from day
-2
onward. By day
0,
positive anomalies associated with this
feature extend from the surface to above
100
mbar. Vertical tilts are rela-
tively small, although there is some indication
of
eastward tilts with increas-
ing height up to the tropopause level, suggesting the possibility of downward
energy propagation from
this
level.
For both positive and negative
cases,
then, the strongest anomalies occur
in the upper troposphere throughout development. At midlatitudes, there
are also substantial low-level anomalies that are initially displaced eastward
relative to the upper level centers; this relative displacement generally de-
creases with time
as
the anomalies approach maximum amplitudes. In con-
trast, the subtropics display little evidence of substantial surface anomalies
prior to development; vertical tilts, if any, appear to
be
predominantly east-
ward with increasing height. These structures are quite distinct from the
vertical structures of disturbances forced
by
local anomalous heat sources as
obtained in simple models
of
the stationary wave response to thermal forcing
(Hoskins and Karoly,
198 l),
suggesting that local diabatic heating anomalies
are unlikely to
be
the proximate source for their development.
The evolving jet structure located over eastern Asia and the western Pa-
cific prior to development
also
appears to
be
associated with pronounced
thermal anomalies. This is illustrated for the negative
cases
in Fig.
I
1,
which
LIFE CYCLES
OF
PERSISTENT ANOMALIES
53
P
P
P
100
250
500
1000
120
E
140
160
180
IS0
140
W
120
b
b'
120
E
140
160
180
160
140
w
la
d
P
I20
E
I40
160
180
Ibo
I40
w
IP
C'
I20
E
140
160
180
160
140
W
120
d'
FIG.
10.
Composite longitude-pressure
(p,
millibars)
cross
sections
of
unfiltered anomalies
(inmeters)forthePACnegativeanomalycasesat45"Nfordays(a)-3,(b)-2,
(c)-
l,and(d)
0.
Corresponding
cross
sections
constructed
instead
at
20"N
are
presented
in
a'-d'.